how to extract highly correlated genes in interaction network ?

Question

[Azhar]

Hi I used lncRNA and mRNA 62k file ,I so file is so big i used the WGCNA packge to get correlation genes, but my TOM obejct file is 24.7GB is so big, How can I extract the highly correlated genes from this file .kindly suggest me some way or script

Thanks in Advance

Answer 1

[Sean Davis]

You may not need the WGCNA package at all if you simply want to get correlations. Take a look at the cor function in R.

That said, no matter what you end up doing, you will definitely want to filter your genes before doing the correlation. Only about 50% of the genes are likely expressed in each sample and some genes will not vary much, so filtering either based on expression level, variation, or both will reduce the number of genes to something more manageable.

[Azhar]

yes you are right but i dont know how to filter i mean by what parameter I have normalized expresson values. so what should i do how can i filter any method or packge and second many Papers used WGCNA for interaction networ may be because it has better or detail algorithm and third if i use R simple correlation how i can build network form it ?

[Sean Davis]

It sounds like you'll need to learn some R before proceeding. Being able to select subsets of data in R is a skill that will serve you well. You can still use WGCNA after subsetting your data if it suits your needs.